On 6/13/19 4:39 AM, Kirill A. Shutemov wrote:
On Thu, Jun 13, 2019 at 05:56:47AM +0800, Yang Shi wrote:
The later patch would make THP deferred split shrinker memcg aware, but
it needs page->mem_cgroup information in THP destructor, which is called
after mem_cgroup_uncharge() now.
So, move mem_cgroup_uncharge() from __page_cache_release() to compound
page destructor, which is called by both THP and other compound pages
except HugeTLB. And call it in __put_single_page() for single order
page.
If I read the patch correctly, it will change behaviour for pages with
NULL_COMPOUND_DTOR. Have you considered it? Are you sure it will not break
anything?
So far a quick search shows NULL_COMPOUND_DTOR is not used by any type of compound page. The HugeTLB code sets destructor to NULL_COMPOUND_DTOR when freeing hugetlb pages via hugetlb specific destructor.
The prep_new_page() would call prep_compound_page() if __GFP_COMP is used, which sets dtor to COMPOUND_PAGE_DTOR by default. Just hugetlb and THP set their specific dtors.
And, it looks __put_compound_page() doesn't check if dtor is NULL or not at all.